Chapter 22 — Key Takeaways

A scannable one-page card. For the full argument and the worked examples, see index.md.

The core claims

• Fire is a physical process, and you cannot read a scene until you understand it. Fire needs the three legs of the fire triangle (fuel, oxidizer, heat); it spreads by conduction, convection, and radiation, mostly upward and outward. Modern furnishings are essentially solid petroleum, with an enormous heat release rate — so an "unusually hot, fast" fire is normal, not proof of an accelerant.
• Flashover is the idea that breaks the old folklore. Past a critical ceiling-gas temperature (roughly $500\text{–}600^{\circ}\text{C}$), radiant heat ignites every exposed surface in a room at once. The fully developed fire then burns everything — including the floor, low corners, and undersides of furniture — producing the low, irregular, "pour-shaped" patterns the folklore mistook for poured accelerant. Flashover artifacts ≠ accelerant evidence.
• Origin and cause is determined by the scientific method under NFPA 921: form hypotheses about where and how the fire started, test them against all the data, and classify the cause as accidental, natural, incendiary, or — the honest answer for a destroyed scene — undetermined.
• The debunked indicators prove nothing about arson. Crazed glass = rapid cooling from the fire hose (an artifact of putting the fire out). Alligatoring (shiny char blisters) = wood burned; nothing about speed, temperature, or accelerant. "Pour patterns" / low burns = routine in post-flashover fires with no liquid present. Each treats a class feature (this floor burned) as an individual signature (someone poured gasoline here) — the Chapter 1 error in a new costume.
• Negative corpus is a fallacy, not a finding. "I found no accidental cause, therefore arson" is invalid, because (a) the fire may have destroyed the accidental cause and (b) the honest classification is then undetermined. An incendiary finding must rest on affirmative evidence of a set fire.
• Accelerants are confirmed in the lab, not by the eye. Ignitable-liquid residue (ILR) is collected vapor-tight (clean metal can or nylon bag — never plastic), with a comparison sample, and confirmed by GC-MS (Chapter 23). It is strong when present; absence does not prove no accelerant (a hot fire can consume it); and a confirmed liquid is not proof of arson without context.
• Bad fire science kills. Cameron Todd Willingham was executed by Texas (2004) on the debunked indicators; Han Tak Lee lost ~25 years to the same folklore before release. Junk science needs no malice — only sincere experts, an untested method, and a courtroom that mistakes confidence for validity.

The method-validity verdict (NAS 2009 / PCAST 2016)

Fire investigation spans the validity spectrum; which fire investigation you have decides everything.

Where they sit: the debunked indicators are at the discredited end with bite marks; modern fire-dynamics work plus instrumentally confirmed ILR is near the top, with toxicology and DNA. The field earned its higher standing not by getting more confident but by getting tested — and by learning to say "undetermined."

What you can honestly say on the stand

• Origin and cause (valid): "Applying the scientific method as set out in NFPA 921, the origin is [X]. The cause is classified as incendiary, based on affirmative evidence — multiple unconnected points of origin and a confirmed ignitable-liquid pattern — not on the discredited visual indicators." Or, when the scene does not permit it: "The cause is undetermined."
• The chemist (ILR): "The chemical pattern recovered from this debris is consistent with gasoline."
• What you must NOT say: "the crazed glass / alligatoring / pour pattern proves an accelerant"; "I found no accidental cause, so it must be arson"; "I know an arson when I see one"; or any conflation of the chemist's validated "it's gasoline" with the scene claim "it was deliberately poured to set the fire."

Key terms (one line each)

• Fire investigation — determining a fire's origin, cause, and development by systematic, scientific-method-based examination; governed by NFPA 921.
• Origin and cause — where a fire began and what started it; cause depends on first finding the origin.
• Accelerant — a substance used to start or speed a fire (usually an ignitable liquid); must be lab-confirmed and is not itself proof of arson.
• Ignitable-liquid residue (ILR) — the chemical remnants of an ignitable liquid in fire debris, confirmed by GC-MS; strong when present, but absence does not prove none was used.
• Flashover — near-simultaneous ignition of all exposed surfaces in a room; produces "pour-shaped" floor patterns in ordinary fires, dissolving the old indicators.
• Negative corpus — the rejected reasoning that infers arson purely by eliminating accidental causes; honest answer is "undetermined."
• Debunked fire indicators — once-trusted visual "arson signs" (crazed glass, alligatoring, pour patterns) shown unreliable by research.

The cold-case line

The Mill Creek fire is incendiary — arson, on valid grounds: multiple origins + an ignitable-liquid (gasoline) pattern, with GC-MS confirmation pending (Chapter 23). It is explicitly not built on crazed glass, alligatoring, "pour patterns" read by eye, or negative corpus — the folklore that executed Willingham. The finding is honest because of how it was reached. It establishes a deliberate fire, not an arsonist: no name attaches to the accelerant.

The themes this chapter advanced

• The validity spectrum (Theme 2) — the central thread: modern fire dynamics + confirmed ILR near the top, the debunked indicators at the discredited bottom; the same discipline at opposite ends.
• The CSI effect cutting both ways (Theme 4) — Willingham's jury over-trusted confident fire testimony that looked like science without being it; the indicators' false certainty is the lever junk science pulls.
• (Also advanced: exclusion over proof — "undetermined" as the honest classification, absence-of-ILR asymmetry; and cognitive bias — ubiquitous indicators as perfect confirmation fodder, §22.3.)